You are here:

  • Print

released: 28 Aug 2007 author: Karmel R & Rosman D

In a feasibility study carried out in 2001 and 2002, the Institute developed a strategy for linking the hospital morbidity and residential aged care databases using date of birth, sex, region of usual residence and event dates. Doubts concerning the efficacy of the linkage strategy were raised at the time because of the lack of either a name or a common person identifier on the two datasets. This paper refines the event-based strategy and confirms its utility by comparing it directly with a name-based linkage strategy.

ISSN 1833 1238; ISBN 978 1 74024 693 4; Cat. no. CSI 3; 150pp.; $30.00

printed copy

Full publication

Publication table of contents

  • Preliminary material (156KB PDF)
    • Half title and verso pages
    • Title and verso pages
    • Contents
    • Acknowledgments
    • Symbols in tables
    • Abbreviations
    • Key points
    • Summary
      • Background
      • Context
      • The linkage strategies
      • Comparing linkage strategies
    • Sections (381KB PDF)
      1. Background
        1. Report structure
      2. Types of transitions
      3. Data
        1. Hospital data
        2. Residential aged care data
      4. N linkage strategy
        1. Linkage protocol
        2. Linkage process
      5. E linkage strategies
        1. Linkage protocol
        2. Linkage process: constrained E matching
        3. Linkage process: basic E matching
      6. Initial comparisons and refining the constrained E linkage strategies
        1. Methods
        2. Constrained matching within SLA group (CSLA)
        3. Constrained matching within postcode (CPC)
        4. Basic matching within SLA group (BSESLA)
        5. Summary of E linking
      7. Match strategy: efficiency comparisons
        1. Links to RAC hospital leave
        2. Positive predictive value
        3. Sensitivity
        4. Summary
      8. Match strategy: analysis comparisons
        1. Examination of distributional differences
        2. Examination of analytical differences
        3. Summary
      9. Conclusions
    • End matter (262KB PDF)
      • Appendixes
        1. Match rules for N linkage and constrained E linkage strategies
          1. Match rules for N linkage
          2. Match rules for constrained E linkage
        2. Illustrating event matching forconstrained strategies
          1. Partitioning the datasets for constrained E matching
          2. Matching RAC hospital leave
          3. Matching RAC admissions
          4. Matching RAC social leave
        3. Preliminary CSLA analysis
        4. Analysis of CSLA missed and false links
          1. CSLA missed links
          2. CSLA false links
        5. Analysis of population size of regions defined variously in terms of postcode
          1. Complete postcodes (4-digit)
          2. First 3 digits of postcode
          3. First 2 digits of postcode
          4. SLA group
          5. Conclusion
        6. Additional linkage comparison tables
      • References
      • List of tables
      • List of figures

Notes and corrections

The current version of the publication is presented above.

Previous versions of files that have been updated or corrected are presented below.

  1. (28 August 2007) Corrected

Recommended citation

Karmel R & Rosman D 2007. Comparing name-based and event-based strategies for data linkage: a study linking hospital and residential aged care data for Western Australia. Cat. no. CSI 3. Canberra: AIHW.